The research also sought to determine why some youths start to offend and what influences others to sustain a noncriminal lifestyle. The survey gathered information from a national random sample of 1,721 young people, plus a booster sample of 808 minority youths. The second stage consisted of life-history interviews with young people who had offended in the past but not within the previous year. Results revealed that half of the males and one-third of the females had committed some offenses, but the majority had committed only one or two minor offenses. Property offending was far more common than violent offending. Half of the males and one-third of the females had used drugs, especially marijuana. The peak age for starting to offend and smoke marijuana was 15 for both males and females. This was 1 year later than the peak age for onset of truancy and running away from home and 1 year earlier than the most common age at which youth began to take drugs other than marijuana. Socioeconomic variables and family size were unrelated to offending, but youths living with both natural parents were less likely to offend than those living with one parent or in a stepfamily, due to less parental supervision, a greater likelihood of a poor relationship with at least one parent, and greater poverty. Females tended to desist from crime abruptly as they left home, left school, formed stable partnerships, and had children. For males, however, desisting was more gradual and intermittent; the positive effects of personal and social developments tended to be outweighed by the more powerful influences of the peer group. Tables, footnotes, appended methodological information, list of other publications from the same organization, and 95 references